Multi-session orchestration for Claude Code with voice control
npx claudepluginhub varie-ai/workstationMulti-session orchestration with voice control. Manage multiple Claude Code sessions, track work progress with checkpoints, and dispatch commands using Apple Speech or LLM-powered voice recognition. Features smart routing, session management, and hands-free workflow control.
Control Claude Code from your phone. An agentic coding orchestrator with AI agent integration, voice control, and multi-session management.
Connect an AI agent like OpenClaw to dispatch commands, approve plans, and monitor progress from Telegram or WhatsApp — while your Mac codes autonomously. Pair with native voice control for a completely hands-free coding experience.
Requirements: macOS 12+, Claude Code CLI
The desktop app bundles the plugin — no separate plugin install needed.
Download from GitHub Releases:
*-arm64.dmg*-x64.dmgOpen the DMG and drag Workstation to Applications.
Launch the app. All Claude Code sessions started from Workstation automatically have the plugin skills available.
macOS Gatekeeper: If macOS blocks the app on first launch:
- macOS 14 and earlier: Right-click the app > Open > click Open
- macOS 15 (Sequoia): System Settings > Privacy & Security > "Open Anyway"
If you prefer to start with the Claude Code plugin (adds skills to standalone Claude Code sessions too):
/plugin marketplace add https://github.com/varie-ai/workstation
/plugin install varie-workstation@varie-workstation
Restart Claude Code. The Workstation app downloads and launches automatically on your next session. To disable: /workstation autolaunch off
git clone https://github.com/varie-ai/workstation.git
cd workstation
npm install
npm run dev
Workstation uses native macOS capabilities that require your explicit approval. macOS will show permission prompts on first use — here's what they are and why they're needed.
| Permission | When prompted | Why it's needed |
|---|---|---|
| Accessibility | First launch | Required for Workstation to manage terminal sessions and capture keyboard shortcuts (voice control hotkey, etc.) |
| Microphone | First voice command | Voice control needs mic access for speech recognition. All audio is processed on-device. |
| Speech Recognition | First voice command | Apple Speech engine requires this system permission. Audio stays on your Mac. |
| Screen Recording | When using full-screen screenshots | Only needed if you want the agent to capture your entire display (not just the Workstation window). Session-mode screenshots use Electron's built-in capture and need no extra permission. |
To manage permissions: System Settings > Privacy & Security > select the permission category > enable/disable Workstation.
Development marketplace for Superpowers core skills library
Harness-native ECC skills, hooks, rules, MCP conventions, and operator workflows
Open Design — local-first design app exposed to coding agents over MCP. Install once with your agent's plugin command and projects/files/skills are reachable through stdio.